A global project delivering an integrated solution for real-time lab data collection and process automation with auto-dosing hardware, built on a microservices infrastructure.
•Developed real-time UI components in Angular using WebSocket/SSE to display live data from IoT sensors, integrating with a pub/sub messaging layer (ActiveMQ, MQTT-style event flow) via Node-RED for device communication.
•Delivered major Angular features such as dynamic REST API-driven forms, interactive weighting sheets, device presets, multi-language (i18n) support, and SSO with RBAC, improving user experience and application security.
•Collaborated in an Agile, globally distributed team comprising frontend/backend developers and PMs, leveraging Jira, Confluence, and Miro to align development with user needs and ensure code quality.
•Contributed to API design, architectural discussions, and cloud deployment during regular Architecture Review Meetings led by architect/lead developer to ensure scalability and performance.
•Successfully released Phase 1 globally to labs and users across three regions (U.S., Germany, China), receiving highly positive user feedback for its intuitive interface, ease of use, and reduced operational workload.

•Independently designed and developed a lab management platform using Angular and Spring Boot to streamline the workflow from experiment placement to reporting.
•Implemented OAuth2 SSO with AWS Cognito federated through Azure AD to deliver a seamless user login experience.
•Initiated Python-based AWS Lambda workflows integrating Microsoft Graph API and S3 to automate experimental data extraction from SharePoint, reducing manual processing by ~8 hours per week.
•Built GitLab CI/CD pipelines with Docker containerisation and Infrastructure as Code (AWS CDK) to automate and standardise deployment across testing and production environments on ECS Fargate.
•Launched in early 2024, the application was adopted by R&D and QA teams across two business entities, reducing manual work and communication overhead equivalent to ~3 FTE annually.

•Established a prototype application using Spring Boot and Angular, integrating R-based Monte Carlo simulation services to help R&D engineers and scientists predict polyurethane characteristics, from recipe setup to results analysis.
•Adopted event-driven architecture with AWS SQS to handle simulation jobs asynchronously, improving system scalability and reliability through retry logic, dead-letter queues, and result aggregation via a polling service to the database.
•Implemented hybrid persistence using MySQL, Redis (NoSQL), and MongoDB (NoSQL) for structured, cached, and dynamic simulation data.
•The MVP was showcased at a company town hall, demonstrating its impact on material evaluation efficiency and generating strong interest from recipe scientists who volunteered to participate in early testing.